Chapter 8
Resistant Desorption Kinetics of Chlorinated Organic Compounds from Contaminated Soil and Sediment 1
2
A m y T. K a n , Wei Chen , and Mason B . Towson
1
Downloaded via UNIV OF CALIFORNIA SANTA BARBARA on June 25, 2018 at 10:08:32 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
1Department of Environmental Science and Engineering, MS-519, Rice University, Houston, T X 77005 2Brown and Caldwell, 1415 Louisiana, Suite 2500, Houston, T X 77002
The chemical release rates from laboratory and field contaminated sediments were studied. Contrary to reports that sorption rates are inversely related to Kow, the slow desorption rates were found to be similar for different compounds. The data were modeled by a two compartment irreversible adsorption and radial diffusion model. Desorption kinetics from the first irreversible compartment can be modeled by radial diffusion and assume an irreversible adsorption constant and soil tortuosity of 4.3. The desorption half life is approximately 2-7 days. Desorption from the second irreversible compartment is very slow, (half-life of approximately 0.32 -8.62 years) presumably caused by entrapment in soil organic matter that increases the constrictivity of the solid phase to chemical diffusion. From the kinetic data, it is deduced that the diffusion pore diameter of the second irreversible compartment is approximately equal to the critical molecular diameter. The mass of chemicals in this highly constrictive irreversible compartment is approximately one fourth of the maximum irreversible, or resistant, compartment. The slow kinetics observed in this study add additional support to the notion that the irreversibly sorbed chemicals are "benign" to the environment.
Particles, or sediments, are the natural depository (and later sources) for contaminants of all kinds including metals, non-metals and neutral organic hydrocarbons. The manner in which contaminants interact with and desorb from particles in natural waters has enormous consequences to all inhabitants of the aquatic
112
© 2001 American Chemical Society Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
113 environment. When hydrophobic chemicals such as petroleum products, industrial wastes and municipal discharges are released into the coastal waterways, the primary fate in every case will generally be to attach to particles in the water column. The manner in which hydrophobic contaminants interact with particles can be either beneficial or detrimental depending upon the specifics of the equilibrium, the kinetics and the mass transport of the adsorption and desorption processes. Mechanistic interpretation of sorption phenomenon is crucial for prior prediction of contaminant fate in the environment. Even though a large amount of mechanistic laboratory research has been directed toward understanding and quantifying the fate of hydrophobic chemicals in contact with naturally occurring particles in the water column, the ability to predict fate and clean-up protocols is still quite poor. The problem is partly due to the complexity of the environment. Different types of soils, chemicals, contamination level, and environmental chemistry could affect the fate of contaminants significantly. In this study, the primary interest is the sorption mechanism of trace level neutral organic hydrocarbons to soil containing a moderate amount of organic matter. Several conceptual models have been proposed in the literature to explain the desorption resistance: (I) The existence of a condensed, glassy, organic polymeric matter as adsorbent. The adsorption to the condensed organic matter phase could be kinetically slow, site specific and non-linear (1,2); (II) The presence of high surface area carbonaceous sorbent, e.g. soots. The adsorption to these high surface area carbonaceous materials could be non-linear, site specific, and limited in capacity (3,4); and (III) The sorbed chemicals are irreversibly trapped in the humic organic matrix following sorption (irreversible adsorption) (5). The term "irreversible" is used to imply that desorption takes place from a molecular environment that is different from the adsorption environment. The irreversible compartment has been found to contain a finite maximum capacity (6). Neutral hydrophobic organic compounds desorb at a similar limit, regardless of the physical/chemical properties of the chemicals. The authors in 1985 (7) first proposed the notion of irreversible sorption as a new mechanism important for hydrocarbon transport. Numerous researchers have shown that the adsorbent surface layer can undergo rearrangement upon adsorption and that this physical-chemical rearrangement can be the cause of non-reversible adsorption (8). These ideas have been tested with a varieties of organic contaminants (5,6,9,10). Recently, Huang and Weber (77) concluded that the entrapment of sorbing molecules within condensed soil organic matter matrices contributes significantly to sorption-desorption hysteresis or sorption irreversibility. Schlebaum et al. (72), presented compelling evidence to show that the anomalous sorption phenomena cannot be explained by the non-linear sorption to a condensed organic phase. Instead, the observations may be attributed to a change in conformation of the humic acid after adsorption, in agreement with the irreversible adsorption model proposed earlier (5). Chiou and Kile (4) proposed that the specific-interaction to high surface area carbonaceous material could reconcile the outstanding features of the nonlinear and the competitive sorption. Interestingly, the data reported by Chiou and Kile (4) can be interpreted as having an apparent saturation capacity of similar magnitude to the irreversible capacity of Kan et al (5,6). Kan et al. (5) proposed that the sorption and desorption of organic chemicals to and from soil consisted of two compartments: (1) a labile desorption compartment, where the chemical can be readily and reversibly desorbed; (2) an entrapped, or irreversible compartment, where the desorption is hindered by soil organic matter due Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
114 to the conformational changes in the soil organic matter following adsorption. The term "irreversible" used herein is in the same manner as is commonly used in the physical-chemical literature, e.g., Adamson (8), Bailey et al. (13), and Burgess et al. (14). A minimum thermodynamic requirement for adsorption and desorption to be irreversible is that there be a physical-chemical rearrangement in the solid phase after adsorption occurs, i.e., the desorption takes place from a different molecular environment than that in which adsorption occurs (8). Desorption from both the labile and irreversible compartments are limited by retarded intraparticle radial diffusion. Chemical desorption from the irreversible compartment shows a constant and different partition coefficient, K Q = 10 L / K g even for chemicals of widely different hydrophobicities. It is proposed that the chemical in the irreversible compartment form an organic "complex" with the organic colloids in the soil organic matter. Therefore, the physical/chemical nature of the adsorbate would be masked by the "complex". A similar conceptualization was proposed by Schulten and Schnitzer (15) based upon molecular modeling. Both the kinetic and thermodynamic aspects of desorption, regardless of the chemical hydrophobicity, are expected to be similar. The possible existence of such a complex has been observed previously (5,16). In this study, the long term desorption kinetics of neutral organic hydrocarbons from field-contaminated sediment (Lake Charles, L A ) were evaluated to further elucidate the mechanisms that control the resistant desorption. s 53
C
Materials and Methods Sediments The Lake Charles sediment was collected near the confluence of Bayou d'Inde and the lower Calcasium River at Lake Charles, Louisiana, by personnel at U.S. Geological Survey in 1995. The sediments were stored in clean plastic zipper bags and chilled to 4°C during shipment. Upon receipt, sediment samples were combined, mixed, and centrifuged at a low temperature to separate the pore water from the sediments, and the sediment was stored in a refrigerator. The sediment contains a large fraction of fine particles. The inorganic fraction of the sediment is mainly quartz, and the organic carbon content is 4.1%. Wet sediment, containing 31.2% moisture, was used exclusively in the experiments. In the following, the reported sediment weight is on dry weight basis. Tenax T A beads (polymeric adsorbent beads, 20/35 mesh, Alltech Associates, Inc., Deerfield IL) were used in desorption experiments. Tenax beads were cleaned before use by successive Soxhlet-extraction with acetone, acetone/hexane (1:1 by volume), methanol, and acetone, respectively, 48 hours for each extraction step. After extensive cleaning, 1 g Tenax beads was extracted with 10-ml acetone. The acetone extract was analyzed by G C / E C D to confirm the cleanliness. Clean Tenax was then baked at 200 C overnight and stored in methanol. Used Tenax was cleaned and regenerated by the same procedure mentioned above. Prior to use, Tenax beads were removed from methanol, baked at 200 C for at least four hours in a clean glass vial. The vial was then sealed and allowed to cool to room temperature. Glassware was cleaned in heated 2% detergent solution (RBS 35 from Pierce Chemical Co., Rockford, IL) overnight at 60 C. Then
Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
115 the glassware was brushed, rinsed with tap water, deionized water, acetone, and oven dried at 200 C. Desorption experiments Five experiments were used to study the release of hydrocarbons for the Lake Charles sediments with the presence of Tenax beads as a sink for the contaminant to desorb. The Tenax desorption procedure is similar to that used in the literature (17,18). In a typical desorption experiment, 0.7 g sediment was suspended in a vial containing 40 ml electrolyte solution (0.1 M NaCl, 0.01 M N a N ) and 0.5 g Tenax. The vial was capped with a Teflon septum and tumbled at 1 rpm for a designated period of time. Tenax was replaced with clean Tenax at 2, 10, 20, and 30 day intervals. A total of 151 days desorption time was monitored. A t the end of each desorption period, the reactor vials were centrifuged for 30 minutes at 1000 g (IEC Centra M P 4 Centrifuge, International Equipment Company, M A ) . The Tenax beads were then carefully taken out with clean stainless steel spatula and transferred to a preweighed 25 ml clear glass vial. Another 0.5 g clean Tenax beads were added into the vial to continue desorption. The contaminated Tenax was extracted with 6 ml acetone in four successive extractions. The acetone extraction was done on a horizontal shaker (Yamato, Model 1290) overnight. C-labeled P C B (2,2'5,5' tetrachlorobiphenyl) or 1,4dibromobenzene were added to Tenax prior to acetone extraction to determine the extraction efficiency. The contaminant concentration in the acetone extract was analyzed by G C / E C D . Over three repetitive extractions, the cumulative recovery of P C B from Tenax was 98%, even though this might not represent the extraction efficiency of the sample since the contact time between sediments and Tenax is much longer than that between the surrogate and Tenax. 3
14
Soxhlet extraction The initial solid phase concentrations of selected chlorinated hydrocarbons on the sediments were determined with the Soxhlet extraction method followed by G C / M S and G C / E C D analysis. Approximately 10 g of wet sediment were mixed with 10 g anhydrous sodium sulfate (baked at 460 °C for four hours), and were refluxed with dichloromethane for 48 hours. 1,2-dibromobenzene, used as a surrogate standard, was added to the sediment before Soxhlet extraction. After extraction, the extract was passed through a 20-mm ID drying column containing about 10 cm of anhydrous sodium sulfate. The extraction flask and the sodium sulfate column were washed with 100 ml methylene chloride to complete the quantitative transfer of chlorinated hydrocarbons and the surrogate standard. The effluent was collected in a KudernaDanish concentrator and concentrated to 2 ml. The concentrated solution was cleaned with a Supelclean™ LC-Florisil SPE tubes (Supelco Inc., Belefont, P A ) and a Nylon Acrodisc 13 syringe filter (0.45 μηι pore size, Gelman Sciences, Ann Arbor, MI). A portion of the final extract was further concentrated and analyzed by G C / M S to identify the compounds in the sediments. The rest of the extract was diluted with isooctane and quantified with G C / E C D (HP Ultra 2 fused silica capillary column, 25 m 1 χ 0.32 mm internal diameter, 0.52 μπι film thickness). The injector temperature was set at 275 C and the detector temperature was set at 390 C. The temperature program was set at an initial temperature of 50 C, for 4 minutes, and increased by 8 C per minute to 280 C and hold at the final temperature for 4 minutes. Equal
Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
116 concentration of 2,5 dibromotoluene was added prior to G C analysis as an internal standard.
Results and Discussion Since Lake Charles sediments were contaminated for over 30 years, the dissolution of the chlorinated compounds should have ceased long ago. Yet, four compounds, i.e., 1,3 dichlorobenzene, 1,4 dichlorobenzene, hexachlorobutadiene, and hexachlorobenzene still exist in significant and measurable quantity. Figure 1 shows the fractional desorption of the four chlorinated hydrocarbons from the sediments by Tenax desorption over 151 days. Even after 151 days desorption, significant fractions of the contaminants in Lake Charles sediments resist desorption. However, the desorption resistant fractions are extractable by Soxhlet extraction. The plot of fraction remaining vs. time showed a progressive decrease in slope, indicating greater and greater resistance to desorption. The dotted lines in Figure 1 were obtained by exponential curve fitting to an empirical two compartment first order kinetic model (Eql) (18). 1
S ^ S o ^ F ^ +F ^
(1)
Where S and S are the mass of sorbed compounds on sediment at time t(d) and at the start of the experiment, respectively; Fi and F (=1-Fi) are the fractions of contaminant presented in the two slow desorbing sediment compartments, respsectively; kj and k ( d ) are the rate constants of the two slow desorbing compartments. For the infinite bath condition, the first order rate constant is related to retarded intraparticle radial diffusion coefficient (D ) by Eq. 2 (19,20). t
0
2
1
2
e
/
c
=
ln2 t
=
1η2
=
2
50%
0.03 R / D
0.693-φ·Ρ„/χ.
( 2 )
2
e
0.03 · R · p · (1 - φ) · OC · K s
o c
2
Where D is the effective diffusivity, D (cm /sec) is the molecular diffusivity of the molecule, χ is the effective tortuosity, p is the solid density (g/cm), k is the first order rate constant (sec" ), φ is the intraparticle porosity and R is the particle radius (cm), O C is the fractional organic carbon content, and Koc (cmVg) is the organic carbon based partition coefficient. The effective tortuosity ( χ = χ/ K ) is the ratio of soil tortuosity factor ( χ ) and a constrictivity factor (K )(19). Theoretically and experimentally proposed tortuosity in unconsolidated material generally ranges between 1.3-3 (19). Constrictivity is a measure of the relative size of the molecule to the pore size. Satterfield et al. (21) proposed the following correlation (Eq. 3). e
w
β
s
1
β
r
r
K =0.98exp(-4.6À) r
(3)
where λ is the ratio of critical molecular diameter to pore diameter. As shown in Table 1, the first kinetic rate constants (kj) are between 0.093 0.309 d" for the four compounds in Lake Charles sediments. The fractional mass of chemicals in the first compartment (Fi) ranges from 0.7-37% for the two 1
Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
117
i 6
i
O
ϋ O
•s
150
0
50
Desorption Time (d) Figure 1. Plots of the fractional mass offour contaminants on Lake Charles sediment versus desorption time (days) following desorption by Tenax beads. The four predominant chemicals in Lake Charles sediment are (a) 1,3-dichlorobenzene, (b) 1,4dichlorobenzene, (c) hexachlorobutadiene, and (d) hexachlorobenzene.
Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
118 dichlorobenzenes and hexachlorobutadiene, but is negligible for hexaehlorobenzene. The second kinetic rate constants (k ) are between 0.00022-0.0026 d" for the four compounds in Lake Charles sediments. The two kinetic rate constants for the three chlorinated benzenes are very similar regardless of the fact that their aqueous solubilities differ by four orders of magnitude. The desorption rates for hexachlorobutadiene are lower by a factor of 2 - 10 than that of the chlorinated benzenes. The significance of these differences between the desorption of the three chlorinated benzenes and hexachlorobutadienes is not known at this time. The reported two-compartment desorption rates are consistent with the previously reported irreversible adsorption data (10). Fu et a l (10) reported a desorption rate of 0.113 d and 0.0016 d" for naphthalene desorption into water from a laboratory contaminated irreversibly adsorbed sediment. It is proposed that there existed two kinetically different regimes in the irreversible compartment. Previous work (5,22-24) demonstrated that the adsorption and desorption to the labile compartment can be modeled by Eq. 1 and K c / K w relation. Since the Lake Charles sediment has been weathered and aged for over 30 years, it is proposed that the kinetic rates correspond to the rates of chemical release from the entrapped and irreversible compartment only, which will be further justified below. Kan et al. (5) observed that most irreversibly adsorbed organic compounds had similar sorption and desorption equilibrium properties ( K Q = 10 ml / g). The effective tortuosity (%) can be calculated from Eq. 1 with the experimentally 1
2
1
1
0
0
55 3
C
e
measured rate (kj) and the same soil parameters (p =2.65, φ = 0.05, R = 50 μιη, K ~ 10 ). The calculated χ are between 2.5 -8.6 with a geometric mean of 4.3. The effective tortuosity is similar to the soil tortuosity estimated previously for soil and other unconsolidated materials (19,23). Therefore, it is proposed that chemical desorption from the first irreversible compartment can be modeled with simple intraparticle radial diffusion of an irreversibly sorbed compound, i.e., the constrictivity factor, K = 1.0. The desorption half life from the first irreversible compartment is between 2 and 7 days. The desorption rate (k ) from die second irreversible compartment is two to three orders of magnitude slower than the first irreversible compartment desorption rate (kj). In Table 1, k values of this study are compared with five other literature reported desorption rate data(70,18,25-27). The chemicals listed in Table 1 include both polar (atrazine and metolachlor) and nonpolar compounds; the range of Kow values varies by four orders of magnitude. Soil organic carbon contents range from 0.26-7.02%. However, the reported desorption rate constants in Table 1 varied by less than a factor of 3 (except for hexachlorobutadiene). It is proposed that a fraction of the irreversibly adsorbed chemical is trapped in the soil organic matter and that the diffusion from the entrapped compartment is limited by a constrictivity factor. In Table 1 are listed the effective tortuosities (χ ) calculated from the desorption rate (k ) and the irreversible isotherm K J J . = 1 0 . The differences in the geosorbents may contribute to the large variation in χ . The geometric mean of the effective tortuosities (%) is 458 for the eleven compounds, which is within the range of effective tortuosity reported by Ball and Roberts for Borden soil (19). Using the soil tortuosity from compartment 1 data, χ = 4.3the mean effective tortuosity of the second compartment (χ* Table 1, Column 8) corresponds to a constrictivity factor, K = 0.01. According to Eq. 2, a constrictivity factor K ~ 0.01 corresponds to λ « 1, i.e. s
0 c
553
β
r
2
2
β
2
553
β
e
r
r
Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
Lipnick et al.; Persistent, Bioaccumulative, and Toxic Chemicals I ACS Symposium Series; American Chemical Society: Washington, DC, 2000.
Lake Carnegie 2.6% OC
Merrimac find sandy loam 2.6% OC Lake Ketelmeer 7.02% O C Hudson River 1-5% OC Lula fine sand 0.27% O C
Aroclor 1242
Atrazine Metolachlor
b
a
Field contaminated, XAD-4 desorption Laboratory contaminated sediment Batch water desorption
Laboratory contaminated sediment Batch water desorption Field contaminated, Column water desorption Field contaminated, Tenax desorption
Field contaminated, Tenax desorption s5
256
53
10
336
6.4
ίο -
635
6.25
10
1 0
10 -
10 -"
s
490
338
338
10 10 10 10 4
y
0.113
NA
458
0.0021
o c
Calculated from Eq. 1 with the first order rate constant (to theright)and K =
553
= 10 ml/g-OC, φ = 0.05, R= 50 μηι.
2
2382
0.0016
549
0.005
NA
10
27
18
75 125
0.005 0.003 NA
NA
26
194 615 0.006 0.0019
NA
0.040 0.038
25
This study
Ref
351
0.003
NA
NA
0.101
2
383 0.0026 (0.68) 0.001 (0.87) 997 0.00022 (0.63) 3634 641 0.0021 (0.99)
(d'r
Rate constant, k
3.51 4.35 8.60 2.49
ab
0.284 (0.32) 0.229 (0.14) 0.093 (0.37) 0.309 (0.01)
/Ce
Number is parenthases are the fractions of contaminant presented in the slow desorbing sediment compartments (F, and F of Eq 1).
Geometric mean
Naphthalene
PCBs
2,2\3,4\5'-PCB 2,2\4,4\5'-PCB
Lake Charles 4.1% OC
(
